The growth of the computerized networks such as the Internet has brought about an explosion in the amount of information available to users. One of the challenges with this wealth of information is how to find and access relevant information easily. Various search engines have been developed to meet this need, such as yahoo.com, altavista.com, infoseek.com, askjeeves.com and google.com. To search for information on the Internet, a user goes to the web site for one of the search engine and enters one or more words known as search terms. The search engine then uses the search terms to generate lists of web sites and/or web pages that best match the search terms. Various indexing and cataloging techniques can be used by the search engine to accomplish the search, such as list based searches, spider based searches or natural language based searches. More recently, meta search engines have been developed that allow a user to enter a single set of search terms at a web site and retrieve a single set of filtered results that are combined from matches found by multiple different primary search engines. Examples of these meta search engines include dogpile.com, chubba.com, thunderstone.com, ixquick.com and metacrawler.com. For a more detailed explanation how search engines work and a comparison of primary search engines and meta search engines, refer to lib.berkeley.edu/TeachingLib/Guides/Internet/FindInfo.html#Outline, cui.unige.ch/meta-index.html, and searchenginewatch.com. Examples of how search engines can be implemented and return combined results are described in U.S. Pat. Nos. 5,864,845, 5,873,080 and 5,974,409.
While meta search engines increase the available resources a user can review, the problem is how to present the results in a useable format. Most meta search engines filter the search results from different primary search engines and present these results according to a combined format unique to that meta search engine. This filtering process can be helpful in that relative rankings can be generated, duplicate results can be eliminated and the amount of information downloaded to the user is reduced, but it precludes a user from reviewing the original content of the results of the primary search engines.
Some meta search engines, such as altaseek.com and 123seek.com, present their results by concatenating the first page of information returned from each primary search engine. In this way, a user can view the first page of the original information from each primary search engine by scrolling down through a single browser window containing the concatenated results. Unfortunately, when a large number of primary search engines are accessed, it is cumbersome to scroll through the large amount of downloaded information in a single browser window. In addition, when a user wants to access information from a particular primary search engine beyond the first page of original information, a new browser window is typically opened and the ability of the user to navigate by backtracking, for example, to previously viewed sites from the original browser window is not possible.
Other meta search engines, such as gogettem.com and searchspaniel.com present their results by opening up a new browser window for each of the different primary search engines. Although this format of presentation is sometimes more manageable than the concatenated scrolling format, the opening of a new browser window consumes additional computer resources of the user for each new browser that is opened and has the same problem of limiting the ability of the user to navigate by backtracking. When several primary search engines are accessed in this manner, the spontaneous spawning of new browser windows can be quite annoying as the user's cursor is constantly being redirected to the most recently opened window until all of the search results have been returned.
A variation on this theme is used by totalnews.com which also spawns a new browser window each time a user clicks on one of a series of buttons on the home page that identify and link to various third-party news web sites. Originally, the home page for totalnews.com used a framing approach in which the content of the selected third-party news web site corresponding to the button selected by the user was framed in the middle of the totalnews.com home page. To make the content information fit into this frame, the totalnews.com home page sometimes would strip off advertisements and other content from the third-party new web site. Ultimately, copyright and source confusion claims by some of the third-party news web sites forced totalnews.com to abandon this framing approach in favor of the spawning of a new browser window.
Another approach to improving browsing and search capabilities is to replace and/or modify the user's browser program. Such approaches are often referred to as a “thick-client” solution because they require the user or client to download and run a program in addition to the user's browser program. Most web sites download content information formatted in a markup language that can be directly decoded by the user's browser program. This is referred to as “thin -client” because the only program that is required to decode the content information is the user's browser program.
Neoplanet.com is one example of a thick-client replacement browser program that integrates web browser, e-mail, chat, search engine and web directory into a single application. In order to run this program, a user must download a very large file from neoplanet.com and install that program in place of the user's browser program. Once installed, the thick-client replacement browser program provides the user with predetermined link connections that, for example, allow a user to select which of one a predetermined short list of search engines or shopping sites will be searched in response to search terms entered by the user.
X-Portal™ findware from centurionsoft.com is one example of a thick-client add-on program that adds functionality to a user's browser program, rather than replacing the browser program. After a user purchases and downloads the findware program file, the findware program adds a button on the control panel of the user's browser program that automatically initiates a search against a customized set of multiples search engines, and then dynamically combines, displays and reranks the results as response are returned from the multiple search engines. The search results can also be compared and filtered through a reference library maintained by X-Portal. Another example of a thick-client add-on program is the iware personal guide program from iwareinc.com. This add-on must be downloaded by the user, after which it creates an intuitive taskbar in the user's browser program. The intuitive taskbar includes, for example, a search feature that directs search terms entered by a user to preselected search engines and then combines the results of those searches when they return. Examples of add-on or plug-in programs for browsers are also described in U.S. Patent Nos. 5,710,918 and 6,041,326. The concept of having content information from one source persist on an user's browser window instead of switching to display content information from a second source as part of an augmented browser control system and persist control system that are added to a user's conventional browser program are described in U.S. Patent No. 5,983,262.
Two of the challenges for the standard browser user interface that provides one page per window display of content information are the delay in loading multiple pages of content information and the difficulty in navigating through multiple pages of content information. Several solutions to delay problem have been suggested, such as deferred display of content information, browser-controlled complexity levels of content information and anticipatory caching of content information, which are described in U.S. Pat. Nos. 5,918,239, 5,987,466 and 6,016,520. Other solutions for the navigation problem have also been suggested, such as non-linear browsing, graphical node structures representing multiple pages of content information, combined index navigator, tree navigator and search term navigator as part of a single browser, and fast return to a search home page, which are described in U.S. Pat. Nos. 5,835,092, 5,877,766, 6,012,055 and 6,052,716.
An alternative approach to solving these problems has been to use side-by-side viewing or nesting of frames of content information. Examples of thick-client applications that have implemented this approach in the context of multi-media content information, grouping requests to multiple sites of content information at the same time, and formatting and displaying nested documents are described in U.S. Pat. Nos. 5,544,354, 6,012,090 and 6,031,989. A commercial application of side-by-side viewing of content information can be found in the Katiesoft™ web interface from katiesoft.com. The Katiesoft™ web interface partitions the user's screen into four quadrant windows and allows content information from up to four web sites to be simultaneously displayed in this side-by-side quadrant arrangement.
Recently, at least two thick-client replacement browser programs have provided a user interface with a feature referred to as browser tabs that allows these thick-client browser programs to display multiple frames of content information in a cascaded format within a single browser window. The Netcaptor™ program from netcaptor.com and the MultiWeb Viewer™ from multisource.com both utilize this tabbed feature to enable simultaneous submission and retrieval of content information from multiple sites. The Netcaptor™ program also provides for a common search term input area that allow a user to enter one set of search terms that are used to search multiple search engines simultaneously. The results of each search engine are returned to the Netcaptor™ program, which then formats those results to be displayed as multiple frames of content information in a cascaded format.
Although these thick-client replacement browser programs may offer a user better ways of viewing multiple sources of content information, the programs are limited due to the large downloads required to install them and the complications associated with supplanting the user's conventional browser program. In essence, these programs must be as compatible with downloading and displaying content information as the conventional browser programs, in addition to providing a different browser user interface. Because these programs are provided with predetermined choices of search engines and web sites, the choices afforded to the user may be limited and updating these choices involves the problems of downloading another update program. Perhaps most importantly, in contrast to conventional browser programs which are provided free of charge, these programs either cost money or require the user to agree to provide personal information that is used to generate customized advertisements which are displayed by the thick -client replacement browser program.
Another alternative for presenting browser information to user is the decision.control program from visualmining.com that provides a thin-client corporate decision portal. The portal includes a “dashboard like” view of multiple sources of content information that are typically maintained on a corporate intranet. The dashboard allows for categorization and searches of corporate information. The portal program is hosted in a corporate environment where the program operates on a dedicated corporate server and communicates with client computers within the computerized network of the company. Different sources of content information from the computerized network of the company may be displayed on the user's browser program simultaneously in multiple side-by-side frames within the browser window. Sources of content information outside the computerized network in the company are displayed in a conventional manner in the user's browser. The search feature on the dashboard of the portal is specialized to search content information inside the computerized network of the company and returns relevant information in a single window frame. The portal program executing on the server utilizes a split-server architecture having a session server for receiving request from clients and a query server that obtains results in response to these requests and passes them back through the session server to the client. This type of split-server architecture is described, for example, in U.S. Pat. No. 5,873,076. Although such a split-server architecture can be well suited for a corporate environment where security issues and the need for common access to private corporate data justify the additional expense of hosting a portal program on a dedicated server, this option is too expensive and complex for individual users.
While numerous solutions have been proposed, the challenge of providing a simple, yet efficient system for a networked thin-client browser to display content information from multiple sources still exists. Thick-client replacement browser programs and add-on programs can offer a user additional features; however, these features can be difficult to manage and maintain and are not consistent with the overall thin-client architecture of the Internet.